Bone microarchitecture assessed by 3D high-resolution peripheral quantitative computed tomography (HR-pQCT) as predictor of fracture risk in patients with glucocorticoid-induced osteoporosis.

Context: Fracture risk in patients on long-term glucocorticoid (LTGC) is only partially explained by impaired areal bone mineral density (aBMD). HR-pQCT derived bone microstructure and bone strength indices could improve fracture prediction beyond aBMD or FRAX in the general population. Abnormalities in these indices could discriminate rheumatic disease patients on LTGC with and without fragility fracture in cross-sectional studies. Objective: To ascertain whether compromised bone quality and strength at baseline could predict incident fragility fracture (IFF) in rheumatic disease patients on LTGC. Design: Setting and Participants: Multi-centre, longitudinal study of patients with rheumatic diseases on LTGC. Main outcome measurements: Spine radiographs, aBMD and HR-pQCT were done at baseline and repeated after a period of 5 years. The occurrence of IFF was documented. Results: A total of 140 patients completed the study, 47 (33.6 %) of whom developed new fractures. After adjusting for age, patients with IFF had significantly worse baseline trabecular volumetric BMD (vBMD), trabecular bone volume fraction, and estimated bone strength at the tibia (RRs 1.51-1.58). Associations remained significant in multivariate models including total hip aBMD, but were insignificant in regressions including previous fractures or FRAX. There were no differences in change in HR-pQCT parameters after 5 years between groups with or without IFF. The area under curve (AUC) generated from a model comprising age, previous fracture history, and tibial trabecular vBMD at baseline (0.710) was comparable with those of FRAX score (0.679/0.702). Conclusion: Impairment of bone microarchitecture and strength in the trabecular compartment of the tibia could help predict the occurrence of IFF in rheumatic disease patients on LTGC. Translational potential of this study: These results demonstrate the potential of developing a HR-pQCT-IFF prediction model specifically for rheumatic disease patients on LTGC.